# BONNIE

> Documentation site powered by Tome

- [Discord channel](https://go-bonnie.dev/channels/discord) - Configure signed Discord interactions, slash commands, durable conversations, and approval buttons.
- [GitHub channel](https://go-bonnie.dev/channels/github) - Configure a GitHub App, signed webhooks, comment admission, event hooks, and durable issue and PR conversations.
- [HTTP channel](https://go-bonnie.dev/channels/http) - Start, inspect, answer, and control durable runs through the BONNIE v1 HTTP API.
- [NATS channel](https://go-bonnie.dev/channels/nats) - Configure Core NATS and JetStream tasks, worker-routed answers, status queries, cancellation, and broker security.
- [Channels](https://go-bonnie.dev/channels/overview) - Connect durable BONNIE runs to HTTP, chat platforms, and NATS tasks.
- [Slack channel](https://go-bonnie.dev/channels/slack) - Configure signed Slack webhooks, durable threads, human input, and activity messages.
- [Telegram channel](https://go-bonnie.dev/channels/telegram) - Configure Telegram bot webhooks, secret verification, private chats, groups, and forum topics.
- [Core concepts](https://go-bonnie.dev/concepts) - Understand runs, turns, steps, durable replay, human input, addresses, and sandbox state.
- [Agent Trees](https://go-bonnie.dev/guides/agent-trees) - Create, test, and build a BONNIE agent from instructions, tools, skills, and context files.
- [Deployment](https://go-bonnie.dev/guides/deployment) - Build and operate a BONNIE agent with persistent storage, authentication, sandbox controls, and safe recovery.
- [Sandboxes](https://go-bonnie.dev/guides/sandboxes) - Select a BONNIE sandbox, restrict network access, and manage working files without weakening the security boundary.
- [Scheduling](https://go-bonnie.dev/guides/scheduling) - Define durable cron work, inspect occurrences, and trigger BONNIE schedules with explicit retry and authorization rules.
- [Tools and Skills](https://go-bonnie.dev/guides/tools-and-skills) - Add compiled Kit tools and on-demand skills to BONNIE agents with clear execution and approval boundaries.
- [Troubleshooting](https://go-bonnie.dev/guides/troubleshooting) - Diagnose BONNIE startup, sandbox, waiting-run, schedule, and recovery problems without losing saved work.
- [BONNIE](https://go-bonnie.dev/) - Durable agent runs for Go. Keep the conversation after a crash, wait for human input, and connect through HTTP or messaging channels.
- [Installation](https://go-bonnie.dev/installation) - Install the BONNIE CLI, prepare a Linux host, and configure a model provider.
- [Quick start](https://go-bonnie.dev/quick-start) - Create an agent tree, ask for human input, inspect its durable run, and compile a static binary.
- [CLI](https://go-bonnie.dev/reference/cli) - BONNIE commands, exact flags and defaults, compiled-agent flags, and operator inspection commands.
- [Events](https://go-bonnie.dev/reference/events) - Runtime event types, journal cursors, live-only Kit activity, replay limits, and activity logging.
- [Journal](https://go-bonnie.dev/reference/journal) - SQLite storage, journal interfaces, lossless records, atomic steps, legacy import, and safe retention.
- [Options](https://go-bonnie.dev/reference/options) - All BONNIE root options, their defaults, validation rules, and managed callback contracts.
- [Runtime](https://go-bonnie.dev/reference/runtime) - Durable run states, execution APIs, human input, cancellation, completion checks, and limits of crash recovery.
- [Testing](https://go-bonnie.dev/reference/testing) - BONNIE contributor commands, hermetic model tests, process-boundary recovery tests, conformance suites, and CI rules.